Breaking Data Silos and Integrating Data

(To better understand the scenario of this case study, we request you to first check on the background details provided here. )

HWPL was compiling substantial data with most of it in the digital format (especially spreadsheets – Microsoft Excel format). These data were generally being compiled for particular purpose such as client submissions, internal monitoring, management reporting and the like. Some of these data were also cross-referred for companywide reporting. However, most of the compiled data were being organized in individual spreadsheets based on the individual presentation requirements with similar records being duplicated across different files. These data were also not organized or structured to allow data modelling or data analytics at scale. It also led to different data being retained in isolated silos thereby curtailing the opportunities to leverage the information by analyzing the data between different information silos within the same firm.

Roboautal and HWPL’s representatives jointly reviewed each primary data source and compiled the same in manageable data sets. Such primary data sources included “Daily Reports” for first-hand data from site; downstream sub-contractor’s payment claims; suppliers’ invoices for data on expenditure; main contractor’s certification for data on valuation of work done. Check and balances were built-in for periodic verification of errors in data sets, identification of sources of such errors and rectification of the errors itself. Measures were also put into place to address the sources of errors including further training of the concerned personnel and tweaking the data entry points to restrain such errors at source itself.

All reports were tied back (as much as possible) to single primary data sources to ensure data trail and integrity. The relevant data sets were structured and cross-referred to extract critical insights including productivity, cost, earned value parameters. The details were further analysed to identify the probable reasons for cost-over runs. Findings were also transformed into pricing strategies that could be adopted in future tenders of similar scope. The business processes were tweaked to enable past data being continually fed into the benchmark calculations for related decision-making.